INTRODUCTION: Heart rate variability (HRV) has shown sensitivity to the acute stressors experienced by defense personnel. This study examines the suitability of overnight HRV as a repeated measure of allostatic load in defense personnel. METHODS: Daily measures of sleep, cognitive load, and perceived exertion were reported for the 12-wk duration of basic military training (BMT) in 48 recruits. Physical activity, subjective well-being, and HRV were measured weekly. The natural log of the root mean square of successive differences of interbeat intervals (Ln RMSSD) and the Ln RMSSD to interbeat interval ratio (Ln RMSSD:RRi ratio) during predicted slow wave sleep were used for HRV. Physical performance was assessed via the 20-m shuttle run and maximal push-up test in weeks 2 and 8 of BMT with predicted V虈O 2 peak values calculated. RESULTS: Predicted V虈O 2 peak increased from 42.6 +/- 4.5 to 48.0 +/- 2.7 mL.kg -1 .min -1 ( P < 0.001). Ln RMSSD was elevated in weeks 7 and 10, and the Ln RMSSD:RRi ratio was elevated in week 10 above all other weeks ( P < 0.05). An increase in perceived exertion ( F = 9.10, P = 0.003) and subjective fatigue ( F = 6.97, P = 0.009), as well as a reduction in V虈O 2 peak ( F = 7.95, P = 0.009), individually predicted an increase in Ln RMSSD. The best predictive model of Ln RMSSD included perceived exertion ( F = 8.16, P = 0.005), subjective fatigue ( F = 8.49, P = 0.004), the number of awakenings during sleep ( F = 7.79, P = 0.006), and the change in V虈O 2 peak ( F = 19.110, P < 0.001). CONCLUSIONS: HRV was predicted by subjective recruit responses to BMT workloads rather than objective measures of physical activity. Improvements in cardiorespiratory fitness depicted recruits who experienced enough stress to facilitate physiological adaptation, which was reflected by a reduction in HRV during BMT. Monitoring HRV and HRV in relation to interbeat interval length may provide a better tool for determining allostatic load than HRV alone.